Speed adjustment considerations

[RulerofRails]

I took a little time to do some tests on station and service stop times. I put the game in slow mode and then double checked that 1 hour of slow time = 1 month regular time.

The basics are that to load and unload an 8-car train takes about 1 month. Each service stop takes about 1/2 a month. This is the measurable part that should be consistent across all locos. I discounted the time it takes the train to accelerate back to full speed.

I also did a calculation for water consumption. This is a little rough, but basically water drops 1% every 2.3* track miles.

Oil consumption is even rougher at 1% every 6.1* track miles.

Hopefully this helps to determine how many stops are necessary in a given year. From the number of stops and average speed with a full consist, we should be able to get a better guess at the range a loco will travel in a year.

My test track run (532 miles - return trip*) would require 3 water stops and 2 oil stops to keep from running out of water or going below 40% Oil. This isn't a great example though because the distance between cities is longer than 227* miles (water will run out). The map was designed with water reduction events because of the distances between cities. So 2 water and 1 oil stop + 2 station stops = 4 and a half months the train is stopped per trip. That's a little skimpy on service, we are definitely running the Oil way below 40%.

A breakdown means that train is stopped for 1 month.

*****Jan-15 2017, Edit of an Edit: Confusion between in-game track mile and track pieces counts had me confused. All counts marked (*) have been re-/dis-/un-corrected. Now they really correspond to track miles as counted by the game! More info on my goof here.

A little table:

[RulerofRails]

So, I did some more testing. This time on the revenue decrease that an aging station undergoes. This is all about the difference between the value that is displayed in-station and what you receive as soon as that cargo is loaded on a train (before rot factor kicks in).

First, I did a test on Medium level, where revenue started out at 50% higher than displayed value in the first year. For the first 3 years there is a rapid drop in this value. In the fourth year I saw 12% higher. Bythe 16th year both values were about equal. The difference in the 20th year was about 8% lower. This value was then flat (no more price decrease) at least until the end of the test in year 24.

On Expert level, the first year difference between displayed value and what I received was almost 22% "bonus" if you will. Once again there is a large drop in the first 3 cylces. The 4th year figure was already 7% less. Then a slower decrease until it bottoms out at 25% less in the 20th year. (Fast decrease is about 9.7% per year. Slower decrease is about 1.4% per year.)

I was mainly looking for some info on how the displayed in-station prices are translating to train revenue. I didn't control the economic state, and I believe it doesn't matter because economic changes are changing the prices one sees on the price map overlays themselves. That figure is seperate and determined before it goes through this particular modifier/formula. I was using the livestock cargo in the B-era. I didn't see anything in this test to suggest that the passenger appeal modifier is affected by difficulty level or economic state. (The final test for that would be to run a gauntlet of tests on the different difficulty levels.)

One caveat: I didn't test for any difference relating to the age of the train. It's unlikely it would make a difference, but you never know.

In conjunction with this test which I ran over a roughly 5 month journey, I recorded the rot factor on cargo price (livestock here). Previous examples have shown that the game's formulas are pretty simple. Across both tests I was seeing a figure of almost 18% in revenue decay over the course of the journey. I plan to test if this is a simple function of elapsed time.

[Gumboots]

I imagine it would be simply the product of elapsed time and the rot factor. I think the latter is set in the .cty file, but PJay's notes aren't very clear.

000 028 : 4 : float : 0,05 + (delivery time / 20), except for Passengers, Mail, Troops (=0,10 too)

I've just checked several .cty files, and most cargoes aren't set to 0.05 at all. Many are set to 0.10, just like express. Oil, rock and ore are examples. Beats me why those cargoes would be losing revenue at the same rate as express. In my experience, oil, rock and ore tend to last rather well. I've seen plenty of examples that are millions of years old.

The fast rot time for concrete corresponds to a value of 0.55. Livestock is set at 0.45. If you lost 18% over 5 months with livestock, that'd be a shade under 4% per month. Not exactly sure how that relates to the .cty file, but it may be just rounding errors on a monthly rate. So as a starting point I'd be assuming monthly loss of revenue in % is probably close to (.cty value x 10).

May also pay to check revenue loss with pax and oil as an easy comparison. That'll tell you if the .cty value is the relevant one.

[RulerofRails]

Ran a bit more testing on the rot factor. The .cty file number is the right one. Except it doesn't seem to apply to passengers. Passengers returned the highest number in the test (6.7% decay per month at a low price point of 15k) even though their value in the .cty file is only 0.1.

I am seeing that the decay rate isn't quite constant through the ranges. Close enough though. For example Milk is average 5.1% decay per month at a price point of 90k, but down around 40k the decay is less at 4.5% per month. This is with a .cty value of 0.55.

I tested Livestock (.cty value of 0.45) again and was averaging 4.3% at 167k. This dropped to 4% at 140k. At 50k, 3.8%, at 20k 3.5%, at 10k 3.1%.

I only tested Oil (.cty value of 0.1) around 100k and it has an average decay of 0.93% at that price point.

So, it's not exact but generally the result is a bit lower than the .cty value. Close enough for my purposes.

---

I am trying to get an approximate measure of the economic value of speed in the game. My feeling while playing is that speed is often expensive in the 20th century. A cheaper, slower train often seems to be a better choice especially when it typically has lower fuel costs to boot as it is often lighter and smaller.

As a really conservative estimate a steamer will be stopped for 1/3 of a year (see OP, 2x load/unload in stations, 2x Water/Sand stops, 2x Oil stops). I am assuming that spur maintenance is used (otherwise in-line stops will make journey times longer). I am also assuming that there are no stops for passing trains, an unavoidable situation in real life. I would say that 1/2 the year stopped and/or lost while accelerating from stops would be more realistic in the 20th century. Nevertheless, I will use the 1/3 of a year figure here. So the train is active for 8 months.

Using the sensitive Milk cargo with a slightly fictionalized value of 5% reveue decay due to cargo rot per month, we can make some calculations. A 20% faster train will earn higher revenue to the tune of 1% per month. For 8 months that's 8%. A 30% faster train will see revenues at 1.5% higher or 12% over 8 months. A 40% faster train: 16%. At 50% faster, we get 20% higher revenue.

We must also consider the maintenance and purchase costs of the faster train. In almost every case, the faster trains cost more to purchase and then maintain. But we will discount this, to maintain a "best case scenario". If the faster train has a fuel rating one step lower, for example Average for the fast train vs Above Average for the slow train we get a 20% higher fuel cost (see the formula in this post) for the fast train. We will guesttimate the ratio of fuel costs to revenue as 1:2. If we halve the 20% higher fuel costs we are effectively decreasing revenue by 5%. So a 50% faster train would have a "best case scenario" advantage of 15%.

[Gumboots]

Ran a bit more testing on the rot factor. The .cty file number is the right one. Except it doesn't seem to apply to passengers. Passengers returned the highest number in the test (6.7% decay per month at a low price point of 15k) even though their value in the .cty file is only 0.1.

Any cargo would have to return an apparent 6.7% at a price of $15k, because the game rounds off to the nearest $1k and 1/15 is 0.06666 ad infinitum. So don't trust that figure because it's probably misleading.

I am seeing that the decay rate isn't quite constant through the ranges. Close enough though. For example Milk is average 5.1% decay per month at a price point of 90k, but down around 40k the decay is less at 4.5% per month. This is with a .cty value of 0.55.

I'd bet that's rounding error too.

I tested Livestock (.cty value of 0.45) again and was averaging 4.3% at 167k. This dropped to 4% at 140k. At 50k, 3.8%, at 20k 3.5%, at 10k 3.1%.

And that too.

[Gumboots]

We must also consider the maintenance and purchase costs of the faster train. In almost every case, the faster trains cost more to purchase and then maintain. But we will discount this, to maintain a "best case scenario". If the faster train has a fuel rating one step lower, for example Average for the fast train vs Above Average for the slow train we get a 20% higher fuel cost (see the formula in this post) for the fast train. We will guesttimate the ratio of fuel costs to revenue as 1:2. If we halve the 20% higher fuel costs we are effectively decreasing revenue by 5%. So a 50% faster train would have a "best case scenario" advantage of 15%.

You're forgetting that the faster train will cover more distance in a year, and fuel cost depends on weight and distance. So if it's 50% faster and covers 50% more distance in the same time, it will have an annual fuel bill 50% higher just by reason of speed.

Then you have your fractions the wrong way around with the fuel economy rating. The increase from Above Average down to Average is 25%, because you're consuming 5x fuel instead of 4x fuel.

Put those two together and your hypothetical 50% faster loco has a fuel bill 87.5% higher*. Now sure, you could probably knock that back a bit because it will also presumably spend more time stopped due to chewing more water and oil, but that extra time stopped will also cut into your revenue.

Short version: yeah, speed is expensive unless you can match it with good fuel economy. This is another reason why I think dedicated express is pretty much borked at the moment.

*Come to think of it, in my experience running Kriegsloks against Class 01's and (occasionally) Mallards on the Italy map, I have seen fuel figures like that for the faster locos. It's one of the main reasons why I don't usually use them, unless I just feel like using them regardless of cost.

[RulerofRails]

What I did for the rot factor test, was to place the game on slow time so I could count how many 12-hour shifts (half-days) it took for the price to degrade one (k). For example, the time between when a load of passengers first hit 15k to the time price degraded to 14k took 58 12-hour shifts. All counts were done on trains stopped as soon as they started loading. For Milk and Livestock I started off with one (k) increments and then after 10 or so occurrences I switched to 5k and then later to 10k increments. I didn't think that rounding should be a big deal with this method. Do you think so? I can post up the spreadsheet with that data if you want. The only slight disclaimer is that I had the game on very fast and relied on my reaction time to push a key to pause the game when each price decay target was reached. This could allow a small one or possibly two day margin of error.

---

You're right. I got the fuel cost level change the wrong way round. Oops. As far as the faster train traveling further in a year, that's true. But, fuel cost will simply be proportionate to the amount of work the train does. A slow train will still burn as much overall, it will just take it longer. I was trying to consider a fast and a slow train doing the same amount of work. I know that this is incorrect when I am using a rot factor calculation as a percentage of the year, but it wasn't meant to be 100% scientific, just to try to get a feel for the potential benefits of a fast loco.

Come to think of it, the fact that faster engines will do more work means that their maintenance and purchase costs can be higher to get the same economic benefit. Probably not as high as the defaults though. This might be a place to use a value such as 20% faster train equals roughly 20% higher purchase price/maintenance cost.

If the engine stats for fuel economy, acceleration, reliability and passenger appeal are equal including the combined engine and tender weights, a faster train is better because you get less rot factor.

My "vision", if you will, for a re-balance is to get all the locos in a particular class and era within a certain economic performance window (A.) so that there aren't any majorly useless locos, and (B.) to encourage the use of a much more varied roster with those locos that are slightly better at one thing able to be used for their strength in a particular role. The way I figure, the average player wants to use a variety of locos. I don't think the economic advantages of any current loco are indispensable. For the strategy-minded player, "knowing" what the best locos in the game are and always using them isn't as fun as being able to use the full range of locos as you find a use for their individual talents. (Do you think having "death trap" engines is good for repeatable gameplay?)

Drastically fudging the weights promises to be the best way to make any economic adjustment regarding fuel cost a reality. Free weight will be really handy in this endeavor. Especially as a negative value to add fake weight. It will be interesting to see how to adjust the fuel cost ratio amongst the classes. I am thinking that Mixed should have highest fuel cost, Express should be lower (depending on average engine and tender weight), but I suspect the heaviest Freights might need to be lowest of all. (Another idea: have the passenger appeal classes such that they identify the class of the loco seeing each class will have specific levels.)

What you see in comparison with the Kriegslok and Class 01 or Mallard is most likely down to the Kriegslok being a level higher (25% cheaper as you corrected me above) in fuel economy and lightest late era steamer at 115 tons. The Class 01 is 180 tons. The Mallard is 215. Using Poptop stats. Pretty easy to see what's going to be cheaper to run.

Ok, so the problem with express goes like this. Because express is already the lightest cargo, the engine + tender weight is going to factor more into the percentage of fuel cost attributable to the engine itself. With further weight decreases, this proportion will get even greater. This could be called fuel cost per load.

Express generation being random and regulated often via event by the scenario's author one will find that it's not uncommon for good Express trains to run at 4 cars or less. I find that in order to really get passengers moving on a map I must cater to a supply of 3 or even only 2 cars worth. Therefore, the fuel cost per load goes up even more. Using a heavy express steamer will likely spell doom in the 20th century. When re-balancing this can be avoided with better fuel economy ratings and light engine and tender weights. (I think of fuel economy ratings as coarse adjustments and engine and tender weights the fine adjustments.)

[Gumboots]

What I did for the rot factor test, was to place the game on slow time so I could count how many 12-hour shifts (half-days) it took for the price to degrade one (k). For example, the time between when a load of passengers first hit 15k to the time price degraded to 14k took 58 12-hour shifts. All counts were done on trains stopped as soon as they started loading. For Milk and Livestock I started off with one (k) increments and then after 10 or so occurrences I switched to 5k and then later to 10k increments. I didn't think that rounding should be a big deal with this method. Do you think so?

Given how simplistic all the other formulae have proven to be, I really can't see this one applying different rot rates depending on cargo price. That's why I think any apparent variation in rot rate when you change cargo price is probably just rounding error. If it isn't, then for some reason the devs must have devoted a lot more sophistication and code to this one little formula in an obscure corner of the game. It's possible, but I'd find it very surprising.

You're right. I got the fuel cost level change the wrong way round. Oops. As far as the faster train traveling further in a year, that's true. But, fuel cost will simply be proportionate to the amount of work the train does. A slow train will still burn as much overall, it will just take it longer.

Which means it will travel less distance over the course of the scenario, so will burn less fuel purely because of that. It's a pretty major factor, IMO.

If the engine stats for fuel economy, acceleration, reliability and passenger appeal are equal including the combined engine and tender weights, a faster train is better because you get less rot factor.

It will also just plain haul more loads, rot or no rot, so will make more money regardless of rot rate (as long as cargo supply is adequate).

My "vision", if you will, for a re-balance is to get all the locos in a particular class and era within a certain economic performance window (A.) so that there aren't any majorly useless locos, and (B.) to encourage the use of a much more varied roster with those locos that are slightly better at one thing able to be used for their strength in a particular role. The way I figure, the average player wants to use a variety of locos. I don't think the economic advantages of any current loco are indispensable. For the strategy-minded player, "knowing" what the best locos in the game are and always using them isn't as fun as being able to use the full range of locos as you find a use for their individual talents.

I agree with this vision.

Drastically fudging the weights promises to be the best way to make any economic adjustment regarding fuel cost a reality. Free weight will be really handy in this endeavor. Especially as a negative value to add fake weight. It will be interesting to see how to adjust the fuel cost ratio amongst the classes. I am thinking that Mixed should have highest fuel cost, Express should be lower (depending on average engine and tender weight), but I suspect the heaviest Freights might need to be lowest of all. (Another idea: have the passenger appeal classes such that they identify the class of the loco seeing each class will have specific levels.)

I agree we'll end up fudging weights to some extent, although I doubt we'll need drastic or wild variations there. In fact I'd be wondering what we'd done wrong if we needed weights all over the place.

I also can't see how negative free weight is going to be any use. I regard it as being of a theoretical interest only. The only reason you'd use negative free weight is is you wanted to totally cripple a loco's hauling power on flat terrain, and I can't see how that would ever be something you'd need to do. Bear in mind that all free weight does is allow hauling more weight on the flat before speed starts dropping. It's the flat terrain equivalent of pulling power. If you use negative free weight, speed will start dropping drastically as soon as you add any consist at all. Pulling power is completely separate, and only affects hauling up grades.

What you see in comparison with the Kriegslok and Class 01 or Mallard is most likely down to the Kriegslok being a level higher (25% cheaper as you corrected me above) in fuel economy and lightest late era steamer at 115 tons. The Class 01 is 180 tons. The Mallard is 215. Using Poptop stats. Pretty easy to see what's going to be cheaper to run.

It's also a distance thing. They're both a lot faster than the Kriegslok, so cover a lot more distance in a year. They also tend to be given priority over the Kreigsloks, so stop less often, which adds more to the distance travelled.

Express generation being random and regulated often via event by the scenario's author one will find that it's not uncommon for good Express trains to run at 4 cars or less. I find that in order to really get passengers moving on a map I must cater to a supply of 3 or even only 2 cars worth.Therefore, the fuel cost per load goes up even more. Using a heavy express steamer will likely spell doom in the 20th century. When re-balancing this can be avoided with better fuel economy ratings and light engine and tender weights. (I think of fuel economy ratings as coarse adjustments and engine and tender weights the fine adjustments.)

See I'm not the slightest bit interested in running 2 or 3 cars on express. If that's all the game is capable of, you might as well just say dedicated express is permanently borked and stick with running mixed consists, in which case we might as well forget about changing cargo weights and appeal ratings. I can put up with 2 or 3 cars when things are starting up, but with a developed system I'd be wanting the larger express locos to regularly (not invariably) haul 6 or more cars at up to 100 mph or whatever they were capable of, with moderate sized express locos like the Schools hauling 4 or 5 at up to 80 mph. Bear in mind I'm thinking of them having stats that allow them to do this, rather than the current situation where most of them are useless as soon as one pigeon perches on the roof.

Although speaking of milk, that often used to be included in real life express consists, for obvious reasons: high value cargo with a fast rot time. So that's maybe something else we should consider. Keep the cargo weight for milk fairly low, so it can give a profit boost to express trains sometimes.

Anyway I agree that economy ratings for coarse adjustments and weight variations for fine adjustments is the way to go.

[RulerofRails]

It could be that the rot isn't a modifier applied on top of other factors and some other time-based formula is causing the differing results. I didn't do extensive testing, only starting at one fairly high price point and working down from there. As this happened, naturally more and more time was elapsing.

I was thinking of the fast train (if it was 50% faster in terms of actual loads hauled/distance covered) as representing 1.5 slow ones. So, I am assuming there is x amount of cargo that I want to move. With the same fuel economy rating and engine weight the fuel costs for 1 fast train or 1.5 slow ones would be the same if we are considering the figure over one year. Does that makes some sense?

Express generation is so variable among maps. I am not sure the main reason behind reducing the passenger amounts on many maps. Perhaps because there is a time factor between when a connection is made and when express would become worthwhile (depending on if and how well you prime for express traffic). Since the game is economic and relies on a growth curve, by the time (let's say 3 years after a decent-sized network of maybe 15 cities) that passengers would become a well-paying idea a scenario author is probably wanting to implement methods to slow down company growth to avoid a company that makes heaps of money resulting in any further goals being a walk-over.

Perhaps we should try to select a typical map size (including number and spacing of cities) for preliminary passenger testing in an attempt to determine a base level for worthwhile passengers. Obviously, there will always be situations where dedicated express will be useless. I don't have a definite map suggestion yet, but it should be one with a large land-mass at the middle of the map containing fairly even city spacing. I would need to go and look at many maps to come up with an approximate average city count.

[Gumboots]

Yeah that makes sense.

If scenario authors want to reduce company profits, nobbling industry production would be far more effective than playing with pax generation. This would also tend to focus the scenario less towards industry and more towards rail, which is generally not such a bad thing in a train game. It wouldn't need much to have a noticeable effect. My guess is that 10% or 20% off industry production would put the brakes on most companies pretty well, at least on expert level.

Edit: While we're on the subject of rot times, I suppose we should get all of them checked. I know some of the 1.06 rot times made absolutely no sense at all (*ahem*concrete*cough*) and it's possible there is similar silliness hiding in the original cargoes.

Passengers being locked to a rot time that isn't related to the .cty value doesn't surprise me, since AFAICT passenger base price is also locked somewhere inside the .exe, but hopefully all the other cargoes actually use their .cty values.

[RulerofRails]

While we're on the subject of rot times, I suppose we should get all of them checked.

Do you mean just copy the values out of the .cty files, or actually in-game testing?

[Gumboots]

I was mostly just thinking of listing the .cty values to make sure none of them were nuts. I'll go through them if you like. I'm doing a lot of cargo checking lately anyway.

[RulerofRails]

On the topic of documentation, in Pjay's notes for the .cty file he made a list of the rot rates for 1.05 cargoes. Here's a short list of the 1.06 ones.

Code: Select all

0.10	Rock
0.10	Ore
0.10	Ingots
0.10	Crystals
0.10	Ceramics
0.20	Machinery
0.20	Electronics
0.30	Medicine
0.30	Dye
0.55	Concrete

[Gumboots]

Can't see any such list in PJay's notes. There's a list of Cargo ID's and a list labelled ?1, whatever that means, but none of the values in that list correspond with rot time floats.

Anyway I went through them all.

Code: Select all

------------------------
1.05 CARGO ROT TIMES
------------------------
Hex value       Decimal
------------------------
CD CC CC 3D     0.10
9A 99 19 3E     0.15
CD CC 4C 3E     0.20
9A 99 99 3E     0.30
33 33 B3 3E     0.35
66 66 E6 3E     0.45
CD CC 0C 3F     0.55

------------------------
Cargo name      Rot time
------------------------
Ammunition      0.15
Aluminimum      0.10
Alcohol         0.15
Automobiles     0.20
Bauxite         0.10
Cheese          0.30
Chemicals       0.10
Clothing        0.20
Coal            0.10
Coffee          0.15
Corn            0.20
Cotton          0.15
Fertiliser      0.15
Furniture       0.10
Goods           0.10
Grain           0.20
Iron            0.10
Livestock       0.45
Logs            0.15
Lumber          0.10
Mail            0.10 NOTE: Actual value is locked inside the .exe.
Meat            0.30
Milk            0.55
Oil             0.10
Paper           0.10
Passengers      0.10 NOTE: Actual value is locked inside the .exe.
Plastic         0.10
Produce         0.45
Pulpwood        0.15
Rice            0.15
Rubber          0.10
Steel           0.10
Sugar           0.15
Tyres           0.10
Toys            0.20
Troops          0.10 NOTE: Actual value is locked inside the .exe.
Uranium         0.20
Waste           0.35
Weapons         0.15
Wool            0.10

--------------------
1.06 Cargo Rot Times
--------------------
Ceramics        0.10
Concrete        0.55
Crystals        0.10
Dye             0.30
Electronics     0.20
Ingots          0.10
Machinery       0.20
Medicine        0.30
Ore             0.10
Rock            0.10

Generally they seem to make sense. A couple of them might be tweaked a bit. The 1.06 concrete value which is well known to be nuts. I might be inclined to slow cheese down a bit and boost coffee a bit. Fertiliser is pretty stable too, unless you mix it with diesel and add a detonator. Meat could possibly get a bit of a boost. Come to think of it, meat and milk should probably have similar values.

I can't see why rice and grain have different values, since rice is a grain (d'oh). Same deal with cotton and wool. Why should cotton rot 50% faster?

Given the half life of most uranium isotopes I can't see why it doesn't have the minimum rot time. If it was radioactive enough to break down significantly over the length of a common train trip, it'd be too radioactive to handle. The value for waste seems too high as well. The game only recycles metals, paper and plastic, so AFAICT it would make more sense for the waste rot time to reflect the valuable components. They're all set at 0.10.

[RulerofRails]

Well, I copied out the values under his "?1" heading and compared them to yours and the values are identical. At least we know that he got them all right now. I had checked maybe half a dozen to conclude that his was the right data.

PS. You missed Diesel from your list.

Here's a little chart that I managed to fool around enough to make. This shows how the factors are distributed for 1.06.

Rot Factor distribution.jpg (9.94 KiB) Viewed 13162 times

Includes all 51 1.06 cargoes. Passengers, mail and troops included in the 0.55 category.

I haven't tried to change the value for passengers. That would be something to do. It wouldn't make sense to increase freight cargoes above 0.55 if we can't also increase passengers. I doubt anything will happen, but you never know. I would only consider changing a lot of values (for example giving more consumer cargoes 0.3+ values) if we have trouble defining the Mixed traffic class.

Because the fuel formula is tied to distance traveled and not speed, all other things being equal (weight of the engine etc.), this rot factor is the only advantage of a faster engine. So in terms of potential Mixed traffic engines, this is where their extra speed gives a natural advantage over pure freight engines. The questions to be settled are, in relation to the true freight engines (remembering that relationship is the thing that matters most), should Mixed engines pay more for fuel, less, or the same? Also this should come into consideration with weights. It would make sense to have the cargoes with higher rot factors lighter. My initial idea is that Mixed traffic engines should have higher fuel costs than true freight or express to give an incentive to split Express and Freight to some degree. We should attempt to decide an initial fuel cost split for testing at some point.

[Gumboots]

I think all the express values are actually locked into the .exe rather than being taken from the .cty files. That would make sense of their .cty rot time values (all 0.10, which is obviously wrong) and would tie in with the lack of change when I tried to change pax base price in my earlier testing.

I wasn't thinking of setting any freights above 0.55. Just maybe nudging a few of the lower values up or down by 0.10 or whatever makes sense. For instance, I can't see why milk and meat have wildly different values. Both are shipped refrigerated, and both have similar spoilage times. They also often have similar delivery prices in the game.

Rot factor isn't the only advantage of faster locos. You also get more turnaround and therefore higher annual ROI, all else being equal. Express should pay less for fuel anyway, purely by virtue of the lighter consists, even if the economy rating and loco weight is the same. I think it would make sense to generally have mixed traffic locos one step down the scale (ie: worse) than pure freight.

And yes I agree about faster rotting cargoes being lighter.

[RulerofRails]

You are right. I shouldn't have said ONLY. Currently, the faster locos mainly have higher purchase prices. When doing the comparison like 1 fast loco does the work of 1.5 slow locos, I was thinking along the lines that purchase price would even out ROI. Obviously that's a generalization that isn't always true.

The other variable, traffic density is hard to quantify. There is some advantage to a fast loco, but personally I don't believe this makes a significant difference. Doubling the most commonly shared track is a much more effective tool than trying to change traffic density by buying faster locos. This effect will become less relevant with the proposed seperate classes as a fast loco will run into slow traffic more regularily. I believe acceleration setting affects the effect of stopping more than top speed.

I still need to test acceleration more. It would be good to get some sort of perfomance comparison between the different levels. Then this could be compared against this rot time data. While talking about Mixed locos, better acceleration than freighters should be another characteristic.

It wouldn't suprise me if the Milk setting was based on the idea of picking up fresh milk from the farm. Just like hauling bulk Concrete by train, it's a wonder that humans make such blunders. Still, I think it's good to keep the 0.55 setting and make Meat the same.

[Gumboots]

Agreed about the acceleration setting. When testing the Schools beta on that large UK/France map I was using an exceptionally large scale for an RT3 map. I can't recall the exact figures offhand, but I think water consumption had to be knocked back by about 50% just to let a train get from London to Brighton without running out. Contrast that with the default PopTop UK map, where IIRC a loco can get there and back with just enough water.

Anyway, point is that with an Above Average acceleration rating, and with the terrain west of London up the Thames valley being dead flat for ages, the train would still be accelerating long past the point where you would think it should hit top speed. The obvious conclusion here is that any loco with an Above Average acceleration rating is rarely, if ever, going to get near its top speed on flat terrain, even if it is given a clear run. The only time it's likely to get there is down a significant grade.

Next obvious conclusion is that things will be worse for lower-rated locos, so the only locos that are likely to attain top speed on the flat will be ones rated Fast or above. We may want to consider this when revamping stats.

I agree about the effect congestion has on faster locos. In general I think fuel cost is currently more important than speed. It can easily outweigh any loss due to rot factor.

[RulerofRails]

From the manual:

Fuel Economy - The fuel efficiency of this locomotive, compared to all other available engines. Total fuel costs are a
product of a locomotive's fuel economy, the distance it travels, and the weight of the loads it carries.
Acceleration - Measures how fast this locomotive reaches top speed from a standing start.

Sounds like this agrees with testing so far.

I ran a few more tests to see if the rot factor setting meant anything to passengers. I tried a range of values and didn't see any effect. I'm not surprised.

I ran a few more tests on passengers, this time in the 1800s. There are so many variables that it's hard to confirm something 100% without running heaps of sets. The figure I am seeing is 3.6% decline per month around the price point of $20 which I take it might be a fairly normal in-game return. I tested some higher price points and the 120k to 130k decline gave me a 2.7% average. In pratical terms though, on Expert level one wouldn't expect to run any freight at those sort of price points so 3.6% should be a decent guideline.

The results are drastically different to my previous figure which was from after year 2000. So, I ran a new test in 2005. This test from 38k down to 13k returned an average figure of 6.9%. I ran all these tests with both Ultra Cool and Ugly locos and they both decay through the same price point at the same rate.

I don't know if it's worth trying to find the years in which the rate changes, but I am glad to find that the game has some more interesting coding. This partly makes up for the fixed nature of pax. I tested Milk in both the early and late eras and the rates are the same. The takeaway from this is that in the early game some freight has a faster decay rate than express.

I'm going to try to test acceleration again to hopefully get some sort of useful comparison.

[Gumboots]

That's interesting. I suspect it ties in with the game cutting pax production for rail after 1970 (I think) when the Boeing 707 is introduced. If you're curious, I'd be looking around that time for the changeover point. There may also be a change when the default newspaper about increasing automobile use pops up (can't remember what year that is).

It makes sense that other cargoes aren't affected by this.